Clap in an empty tiled bathroom and the sound doesn't stop when your hands do. It hangs, smears and slowly dies. Clap in a room full of sofas and curtains and it is gone almost at once. The difference has a name, reverberation time, and in 1895 a young Harvard physicist worked out exactly what controls it using an organ pipe and a stopwatch.

A sound that won't leave

When you make a sound in a room, the waves hit the walls, floor and ceiling. Each time, a little of the sound is soaked up and the rest bounces back. The bounced sound hits something else, loses a bit more, and so on. What you hear after the source stops is that cloud of bouncing sound slowly draining away.

To compare rooms, acousticians time how long the sound takes to fall by 60 decibels, a drop in intensity by a factor of a million. That time is called RT60. A short RT60 is a "dry" room. A long one is "live" or boomy.

The problem at Harvard

Harvard's Fogg Art Museum opened a new lecture hall in 1895, and it was nearly unusable: speech smeared into mush. The problem landed on a young assistant professor of physics, Wallace Sabine. He took an organ pipe and a stopwatch, sounded the pipe, and timed how long the sound stayed audible after it stopped. Then he changed things: brought in rugs, seat cushions, more people, and measured again.

According to the American Physical Society's history of his work, he found that in the Fogg hall a spoken word stayed audible for about 5.5 seconds, against an ideal he put at about 2.25 seconds. Every added absorber shortened it, and a person's body soaked up about as much as six seat cushions.

Sabine's rule

From many such measurements he found a simple relationship. Reverberation time grows with the room's volume and shrinks with the total absorption of its surfaces:

RT60 ≈ 0.16 × V ÷ A

Here V is the volume in cubic metres, and A is the absorption: each surface's area multiplied by how much sound it soaks up (from 0, a perfect mirror for sound, to 1, an open window). Bare plaster or tile absorbs a few percent. Heavy curtains, carpet and people absorb far more.

The logic is intuitive. A bigger room means sound travels farther between bounces, so it makes fewer hits per second and loses energy more slowly. More absorption means each hit takes a bigger bite. Volume on top, absorption underneath.

Build a room and clap

This model uses a simple box-shaped room, so its surface area follows from its volume. Drag the sliders, then press Clap to watch the sound level fall. The dashed line marks the 60 dB drop, and where the curve crosses it is the RT60. The room sizes are rough and illustrative.

0.0 s

What the number is good for

Different uses want different reverberation. Speech needs it short, so each syllable dies before the next arrives. Music often wants it longer, so notes blend and sustain. That is why a lecture hall and a concert hall are built differently, and why the same orchestra can sound thin in a carpeted rehearsal room and rich in a stone church.

Try holding the volume fixed and dragging absorption up: the room goes from boomy to dead. Then hold absorption fixed and grow the room: it gets livelier again. Sabine's trade-off is the whole craft of room acoustics in miniature: you can only tune a room's echo by changing its size or what its surfaces are made of.

His formula is an approximation. It works best in rooms where sound is spread evenly and absorption is modest, and it overestimates the reverberation time in very dead rooms. But it was the first time anyone could predict a room's sound before it was built, and Sabine went on to advise on the acoustics of Boston's Symphony Hall, which opened in 1900.

The take-away

Next time a room sounds echoey, you can read it: big volume, hard surfaces, nothing to soak the sound up. Hang a curtain or add a rug and you are moving the bottom of Sabine's fraction.